How to share and utilise expertise in a police forensic department through externalisation and mutualisation

Science & Justice xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Science & Justice journal homepage: www.elsevier.com/locate/scijus

How to share and utilise expertise in a police forensic department through externalisation and mutualisation ⁎

Emmanuelle Ernea, , Mauro Cherubinib, Olivier Delémonta a b

Faculty of Law, Criminal Justice and Public Administration, School of Criminal Justice, University of Lausanne, Switzerland Faculty of Business Sutdy, Information Systems Department, University of Lausanne, Switzerland

A R T I C LE I N FO

A B S T R A C T

Keywords: Fire investigation Knowledge management Case library Ethnographic study Computer-supported cooperative work Human factors

The technique of fire investigation is a forensic domain in which expertise and analogies play a central role. To learn how fire investigators use these analogies to support their work, we conducted an ethnographic study in a Swiss forensic police department. To propose a suitable knowledge-management strategy, we also evaluated the knowledge conservation and sharing within the department. Our results highlighted that actionable knowledge is registered mainly in the investigators’ memories of a few, very experienced, individuals. Without experience with fire-incident investigations, an agent generally requires help from a more experienced colleague, who will then use his memory to find a similar case, which can contribute to the solution of the ongoing one. The research also established that knowledge is exchanged orally during on-site investigations and that knowledge receivers are generally those who are present on the scene. Using these findings, we suggest building a case library to support the externalisation and sharing of knowledge.

1. Introduction Forensic science encompasses a wide range of activities and skills involving (crime) scene examination, and sometimes going as far as evidence presentation in court. In this paper, we look at the forensic activity of scene examination that, despite attracting little attention from the academic community, remains the core practice of the overwhelming majority of forensic practitioners. More specifically, we focus on one type of situation for which scene examination is undertaken: fire incidents. Due to this particular type of investigation – the elucidation of the origin and cause of the fire, of its propagation, and, when appropriate, of criminal responsibilities – it can be conducted by specialised units. But usually, it is carried out by police forensic units that are involved in all types of scene examinations. Yet, fire investigations differ by both the context and the environment –frequently almost entirely destroyed – in which they take place. In Switzerland, fire investigations are realised or supervised mainly by investigators with extensive experience in this domain. This highlights that fire investigations are difficult to solve for agents with less experience in the field, hence the experts’ experience is crucial. Consequently, fire-investigation knowledge is highly linked to a restricted number of individuals and the risk of the loss of knowledge and competence, due to retirement or departure, is significant. To avoid such deleterious situations, it is necessary to externalise ⁎

the relevant knowledge and to make it available and actionable to others. Several methods, such as case-based reasoning or a case library, can assist in this perspective, and the choice of the most suitable one is not straightforward. It should rely on a deep understanding of the interactions between the individuals within a forensic unit and of the forms of knowledge exchanges that take place. Therefore, we conducted an ethnographic study in the forensic department of a state police (Commissariat d’Identification Judiciaire, Police Cantonale de Fribourg, Switzerland) in which the main author of this paper is employed part-time. We wanted to learn how, in particular, knowledge gained through fire investigation was maintained, used, and shared. We also looked at unveiling and assessing the roles and mechanisms of using case analogies in the investigation of new fire incidents. 1.1. Fire investigation as a knowledge-based activity In a police forensic-unit context, fire scene investigations might appear largely as case-to-case activities, each fire incident being the start of a new and isolated investigation, conditioned by the characteristics of the incident. In fact, the investigation process is strongly embedded within a continuum of the practices of a unit and of the individuals composing it. To help solve the new situation, past investigations provide valuable knowledge that investigators can use –

Corresponding author at: University of Lausanne, Ecole des Sciences Criminelles, bâtiment Batochime, CH-1015 Lausanne, Switzerland E-mail address: [email protected] (E. Erne).

https://doi.org/10.1016/j.scijus.2019.12.004 Received 8 August 2019; Received in revised form 11 December 2019; Accepted 15 December 2019 1355-0306/ © 2019 The Chartered Society of Forensic Sciences. Published by Elsevier B.V. All rights reserved.

Please cite this article as: Emmanuelle Erne, Mauro Cherubini and Olivier Delémont, Science & Justice, https://doi.org/10.1016/j.scijus.2019.12.004

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

more or less implicitly – by analogy: the fire under investigation is scrutinised as a possible “reoccurrence” of a past incident(s). Past solutions or successful strategies are re-applied to current fire incidents. This approach refers to a concept of recurrence or repetition that plays a pivotal role in many applications in forensic science, in particular, for intelligence purposes. Physical traces (DNA profile, shoe marks, etc.) are commonly used to link offences, revealing the repetitive actions of perpetrators [1]. Other dimensions of repetitions are also taken into account in fire investigations, thus revealing possible problematic situations or informing decision-making: The detection of serial arsons can influence the operational investigation strategies [2]; and evaluating the predominant “modus-operandi” of accidental fire can usefully impact prevention strategies [3]. In Swiss police departments, the repetitive dimension of fire incidents is mainly considered for the identification of serial arsonists and for the mitigation of risk situations. But, the recurrence of similarities between fire incidents is already a basis of the actual practice of firescene investigation. Facing a new incident, investigators rely on the lessons learnt from past cases that could share some degree of similarity with the current incident: similar environment, materials, or heat sources involved, etc. Findings or successful strategies from these past cases are then integrated as actionable knowledge in the investigation process of the actual case. This kind of repetition between cases is central in the investigation process: making analogies is the core of cognition [4], and recognition of common patterns in two cases is an essential skill for an experienced agent, independently from his domain of speciality [5–8]. The context of fire investigations varies substantially in terms of place, appliances involved, and potential criminal dimensions. Damages can hide some clues. This implies the impossibility of using a predefined procedure to guide all investigations. Context analysis and adaptation is essential in these types of case. Consequently, pattern recognition is the main tool of fire investigators and is linked to the individuals’ experiences. This highlights the fundamental role held by the repetition of cases in fire-incident investigations. Existing research suggests that recognizing similarities is an important means of knowledge conservation and sharing. But the nature and the extent of this reasoning through repetitions is not formalised, and its added value still needs to be understood. Despite the common use of similarities in fire investigations, no study has been conducted to better understand which elements are used to establish links between cases, the level of abstraction of these links and how they could differ or not between individuals. This is precisely the purpose of this paper: to report the results of an ethnographic study conducted within a police forensic department. We studied the ways a forensic department manages fire-investigation knowledge and similarities between cases. In consideration of relevant scientific research, these results lay the foundations for the design of a knowledge-management system that can support operational investigations and contribute to perpetuating relevant knowledge. We believe that such research can substantially contribute to understanding how expertise influences the use of repetitions in fire investigations, either to solve new cases or to manage knowledge within a forensic department. Hence, this is a prerequisite to the improvement of fire-investigation knowledge management. This research is part of an ongoing project for understanding, structuring, and mutualizing the knowledge generated within a forensic unit during the investigations conducted by each forensic investigator of fire incidents.

support with ongoing investigations. In this perspective, the standing of experienced fire investigators is comparable to that encountered in many other domains, epitomised by the distinction between an expert and a novice. One of the main differences between experts and novices is their problem solving ability: experts will rapidly find the best solution and identify clues that others do not [7,9,10]. This superiority in problem solving can be explained by an improvement in reasoning and cognition gained from experience. Novices rely mainly on formalised (or explicit) knowledge and are fully conscious of every step made within the reflective process [11]. With the accumulation of experience, and consequently of tacit knowledge, their reasoning becomes increasingly intuitive and unconscious [12]. According to Collins and Evans, novices possess “ubiquitous knowledge” [13] like all humans do. As they are increasingly involved in their respective practices, they become members of the investigator community, listen to stories about the field, and acquire its codes and vocabulary [14–17]. This immersion in field work will enable them to develop specialist tacit knowledge, more precisely, “contributory expertise” [13]. In parallel, the accumulation of knowledge and experience gained through practical work alters a person’s memory structure, enabling them to establish easily the similarities of cases [4]. Each time a case is solved, the investigator will commit to memory the details by classifying them using one or more general concepts by which the case is represented. With the accumulation of expertise, these networks of concepts will became larger, more complex, and distributed on many levels of abstraction [4]. Consequently, similarities drawn between situations will be different for a novice and an expert [18]. Generally, novices will establish similarities at a lower level of abstraction. Whereas, the expert’s complex network will enable him to recognize many similarities at several levels of abstraction [4,18]. As a result, the experts possess “contributory expertise” that enable them to exercise their activity with competence [13]. With such a cognition system, experts are able, in almost any situation, to recognize many similarities between cases and to reuse elements of past cases to help solve the present situation. This also enables experts to make heuristics [4]. These elements help us understand an expert’s analytical superiority but also highlight the fact that such capacities require years of training and remain only with individual experts. As few works of research presented in this literature review relate to the forensic domain, and none specifically to fire investigation, in a first step, we evaluated if expertise induces cognitive and performance differences between investigators. Then we wanted to understand how fire-investigation experts use repetitions and similitudes to solve fire incidents. As these elements are understood through the results of our ethnographic study, we propose a suitable methodology to extract, conserve, and share fire-investigation knowledge. As the environment of the forensic unit is crucial to implementation of the methodology, we first study how knowledge is actually managed in this unit. Then, we define the best knowledge-management strategy that can be applied within this service. The following methodologies are currently used to guarantee companies’ competitivity [19] and are diverse: supervision [11,20], community of practice [21–23], case-based reasoning [24–28] and case library [29–32]. Even if supervision and community of practice (exchange between more and less experienced investigators) are used within almost every forensic department, through interraction between investigators, they are not seen as a knowledge-management tools; and their use is rarely formalised. Case-based reasoning relies on the reuse of past cases and on successful strategies to support the resolution of new cases. Case-based reasoning systems can be quite different in terms of the automation of problem solving and the interraction with the user [33]. One of the simplest forms is a case library in which similar past cases are proposed to the user to support his reasoning.

2. Literature review Within the activity of fire investigation in Switzerland, it is widely acknowledged that experience – in terms of number and diversity of incidents investigated – is an essential component of proficiency. As experienced fire investigators are generally recognised as experts, they are mentors for colleagues and are generally requested to provide 2

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

3. Methodology

Table 1 Summary of the population characteristics.

An ethnographic study was conducted to gain a clear understanding of investigators’ reasoning and of how fire-incident investigations and case knowledge are managed within a given forensic department. We chose this approach because the main subject of this research is humans. Understanding how humans think when running a fire investigation and how knowledge is managed within a particular forensic service requires us to study the agents within their work environment. Ethnographic studies are particularly suitable in this perspective as they enable us to study the fieldwork of a particular group of individuals, in a specific environment, and for a long period of time [34]. Recently, Wyatt and Wilson-Kovacs [35] used this approach to study crime-scene examiners and to better understand their role within the criminaljustice system. One of their conclusions is that “… ethnographic lens has a lot to offer the academic and practitioner literature on crime scene examination” (p. 7). In our study we used participatory observations of forensic investigators and semi-structured interviews. We use these two datacollection methods for their complementarity and compatibility with the daily work of forensic investigators. A similar approach was applied by Köpsén and Nyström [11,20] to study novice education through supervision by experts in a forensic laboratory. As duty constraints in such an operational unit require high flexibility with time management, the participatory observation of the agents seemed to be a very suitable approach for the ethnographic study [34,36]. The unpredictability of fire-incident occurrences and the fact that one of the authors is a fire specialist in this department meant that participatory observation was the best methodology to run an ethnographic study, without having a negative effect on investigations. The observations were conducted both at fire-investigations scenes and during working hours at the office. In some particular situations, we discussed with the investigator after a particular observation in order to better understand their reasoning related to a particular action or to obtain further explanations about a specific statement. The data collected through our observations was completed and corroborated through semi-structured interviews. We prepared an interview plan in order to be systematic and to have a data collection as exhaustive as possible. The interview was based on four dimensions: (1) the professional situation (i.e. education, seniority, role within the service…), (2) the way knowledge gathered through case investigation was conserved by individuals and within the department, (3) the way this knowledge was shared, and (4) how relations between cases were made and utilised. The investigators were asked to provide examples of similarities between cases, if they were able to remember any. We asked for details about the types of cases, the types of similarities and the purpose of having established such connexions. Interviews were conducted, when permission was given, in the investigator’s office and recorded after agreement. Our ethnographic study was conducted between January and July 2017. During our observations, we paid special attention to the individuals involved, the environment of the forensic unit, the tools used and some particular reactions. Notes were sometimes taken during our observations. As the main author plays an active role in the fire investigations, notes were mostly taken after the completion of the examination of the scene.

Pseudo

Seniority (years)

Number of cases treated

Robin André Rachel Cindy Marco Aurélie Laura Julien

1 3 5 8 14 15 15 24

10 13 ? 23–30 50–100 50–100 50 400–500

3.2. Population The observations were conducted on a regular basis. The total workforce of the police forensic department during the studied period was 17 investigators: 8 women and 9 men. 10 of them had an academic degree (MSc in Forensic science) and they had between 1 and 24 years of experience. This department possesses a fire- investigation specialist team that is composed of 3 people: an experienced investigator (Julien), a team manager, and fire-investigation specialist (the main author of this article). The interviews were conducted with 8 of these 17 investigators. They were selected to be representative of the larger set of investigators in the forensic department. The interviews lasted one hour, to have sufficient time l and to not interfere with their work. The demographic details of the interviewees are presented in Table 1. There were 4 men and 4 women of different hierarchical levels, with between 1 and 24 years of forensic experience (mean: 10.63 years). 6 of them had an academic degree, all in forensic science (MSc level). The 2 others completed another professional education before entering the police academy and served for several years as police officers, before joining the forensic department where they underwent forensic training. All had received a theoretical training in fire investigation (either at university or during continuous education) and completed their education in the field, under supervision of skilled investigators. At the university, the theoretical training consists of tow modules of 28 h, each completed with a practical training of 48 h. During continuous education, it consists of one basic module of 6 days and is completed with specialised modules of 3 days. All of the investigators were, specialists in scene investigation, but some also had specific roles within the department. In particular, one of them – Julien1 – was considered a fire-investigation specialist due to his expertise and the high number of fire incidents he investigated. The amount of fire investigations conducted by these investigators during their career varied from 10 to about 500. 3.3. Research location The ethnographic study was conducted directly in the police station within the forensic department and at the different fire scenes in the canton of Fribourg. 3.4. Information systems in use in the forensic department At the office, two databases are available for managing the information related to fire incidents. The first one is the general police information system (Zephyr) and is accessible to every member of the police. It contains all the information on cases and all documents produced during police activities, including pictures and reports. The second one is the LIMS database specifically designed for the forensic department. It enables the police to manage forensic evidences and data related to cases handled by the forensic department. It

3.1. Validation We presented our results to the participants during a meeting in which they had the opportunity to make comments. All the data were validated as presented, which implies that the implication of the main author within this department did not induced bias on the interpretation of the results.

1

3

All names are fictitious to preserve anonymity.

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

a formal group of specialists. As in other domains, experts in fire investigation are individuals who possess an expert ability in problem solving. This characteristic is recognised by other forensic investigators.

includes all the traces and exhibits collected within a case, as well as the results of the analyses performed, thus providing a global overview of each case. The current architecture of the database was implemented in January 2016. This tool was developed because Zephyr was not designed for the management of forensic related information. Concerning fire- incident investigations, LIMS contains mainly the summary of the investigations with the findings, and the evidence collected at the scene for further laboratory analysis. The pictures and reports are stored separately, but direct access to all information needed is possible through direct links in the LIMS.

4.2. The use of analogy and case repetitions During this study, we find that case plays a significant role in the problem-solving of fire incident investigations. Facing a new case to solve, investigators generally compare the new case to situations they have previously investigated, either already during the scene examination or later during discussions at the office. This was particularly evident with Marco and Julien, two of the most skilled investigators, who regularly and explicitly refer to old cases to explain their reasoning about ongoing investigations. They explain the similarities or differences between cases and the way this supports (or not) their hypotheses of the origin and cause of the fire. If the discussion takes place at the office, they sometimes rely on pictures from old cases to illustrate their reasoning and to show how the elements are similar or not (see Fig. 1). At the scene, as they do not have pictures, they describe how traces or elements of old cases look like or differ from what they observe. During the interview, Robin (the least experienced investigator) mentioned a case in which Bernard (another experienced member of the fire-investigation specialist team) assisted him. He was facing a fire possibly caused by self-combusting hay (see Fig. 2). Robin said,

4. Results 4.1. Importance of expertise in the investigation process During the interviews, the influence of expertise – either that of the interviewed individual or of others investigators – was discussed. The central role of fire-investigation specialists and how their expertise facilitated solving cases was highlighted. Four inspectors mentioned that they did not feel confident in this type of investigation and generally rely on more experienced investigators to assist them. Aurélie said, “Except for really simple cases, such as kitchen fires, I generally need a second opinion on my cases.2” We find that almost every fire incident was discussed with Julien. Due to his expertise and knowledge, other investigators consider him a reference for fire investigation. Some of the interviewees mentioned that people with extensive expertise such as Julien, reasons in a manner that is difficult to follow. One of them explained that sometimes he was not able to follow Julien’s reasoning in the field and that he finally understood, after a detailed discussion with Julien at the office. Some also reported that more experienced specialists seem to see clues that are not perceived by other investigators. For instance, Laura stated,

“Concerning this fire incident, it was my first situation of self-combusting hay - but we did not see the typical burn patterns of this phenomenon. As it was my first time, I was not sure of the cause, but Bernard directly said that it was this cause due to the typical smell.” This ethnographic study has enabled us to highlight the fundamental role of using similar past situations in order to gather valuable information for the investigation of fire incidents. When an investigator requires someone else’s assistance, the request “Have you ever investigated a similar case?” often appeared in the discussion. Through our observations of the participants, we noticed that the trend to consider previous cases was also significant for less experienced investigators. For example, at the scene of a burned barn where the self-combusting hay was considered, Aurélie said that the smell of the hay was the same that she had smelt on other cases, and she considered it as typical of self-combusting hay. She stated,

“When investigating a scene, a skilled investigator, already has two or three possible causes for the fire in his mind and look for relevant elements determine the cause. If I am alone, I think that I would not go to see those things because I would not be able to know where to search.” And Aurélie added,

“This smell is so particular that even the first time you smell it, you will recognise it and you will never forget it”.

“Julien often explains how things happened. […] so we learn about what we need to look for, but I am not sure that I would be able to redetect such things on the scene because the picture that we saw is out of context and sometimes objects are tiny.”

Another example occurred when Robin was investigating a vehicle fire. He made a search in the LIMS database in order to retrieve several vehicle fire cases and to compare the traces with these cases and the ongoing one. He used this knowledge to evaluate and to balance several hypotheses of the cause of the fire. He made the research by simply using the “type of crime” and “type of place” fields of LIMS to obtain all fire vehicles previously investigated. It retrieved 4 cases (the database was only one year old). The comparison with previous investigations enabled him to confirm that the burn pattern in this case was similar to those where the fire started at the front of the vehicle. These results show that case repetitions are central when reasoning about fire investigations. Table 2 summarises the results obtained after asking investigators if they remember situations in which they recognize similarities between two cases. The first similarity concerned two cases for which more than one possible cause remained plausible after the investigation. Having to write her report, Rachel used this similar previous case as inspiration and guidance to write the report for her case. Concerning the second similarity, the old case was the one where Bernard helped Robin to establish the cause of the fire as self-combusting hay. After having investigated this new fire where hay was involved Robin said,

According to the literature review, the difficulties expressed by Laura and Aurélie about their ability to follow or to copy Julien’s reasoning is probably the consequence of cognition differences. Experts think faster and use features different than those novices do to build their reasoning and to conduct investigations. Marco, despite having a substantial experience in fire investigation, did not consider himself a fire specialist (he was not formally a member of the fire-investigation specialist team). Nevertheless, he pointed out the fact that other investigators had often required his help and knowledge when they had to solve case. He explained, “I do not consider myself a fire specialist. People ask me questions […] also because I investigated many fire incidents and maybe because I know about building materials or can help understand how things happened.” This statement illustrates that actual expertise is perceived as fundamental, and that it does not necessarily correspond with belonging to 2

All citations were freely translated from French. 4

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

one has been already discussed above. The second established a connection between an old case in which the fire was set in a trunk and the new case where it was hypothesised that the origin of the fire was situated in the engine. Here the similarities between the two fire incidents were used in order to exclude a hypothesis. Robin explained, “Concerning those two cases, you can really see the difference in the burn pattern intensity in the engine and it helps to situate the zone where the fire started.” During the whole interview, Robin was really clear about which similarities were recognized, how they were made and how he used this information. Whereas, the results with Julien were a bit different. First, he said that he did not remember situations in which he established connections between cases and, after a few seconds, he cited the two examples presented in Table 2. The first one concerned the geographical proximity between many fires of small public buildings. The second involved two wooden bridges where the fire was set with similar fire-starter blocks. Julien stated, “It was the same fire-starter blocks, so I immediately put the two cases together.” The examples given by Julien are interesting because they are really different from our observations. In almost every fire investigated, he referred to previous cases to support his reasoning. These relations between cases were mainly used to identify the origin and cause of the fires and pertained to the type of appliance, type of defect, and the physico-chemical phenomenon involved. Whereas, the examples he mentioned during the interview related to the possibility of a common arsonist for many fire incidents. This step of reflection generally occurs after the determination of the origin and the cause of the fire. These

Fig. 1. Marco and Julien discussing about a fire investigation at the office.

Fig. 2. Example of traces of hay self-combusting before the fire start.

results highlight the fact that most of Julien’s reasoning is automatic and unconscious. He only remembers the connections established at the final steps of the fire investigation where links are formalised and used to produce intelligence. Our observations and interviews highlight the fact that recognizing similarities is a central reasoning scheme for every investigator. But expertise induces substantial differences in the intuitive part of this

“When this new case occurred, similarities were rapidly seen, we tried to see if the same signs were present and if the time difference between the fire starting and the day the hay was put in the barn was similar.” In this situation, the presence of the hay was the element that enabled the investigator to make a connection between the two cases. The two other cases reported by Robin were vehicle fires. The first 5

Science & Justice xxx (xxxx) xxx–xxx

reasoning.

Report writing guidance Traces and chronology comparison Traces comparison in relation with the cause of the fire Traces comparison depending of the origin of the fire Research of cases that could be part of a serial phenomenon Possibility of a common arsonist Cases with more than one non-excluded hypothesis Hay Vehicle Vehicle Geographical situation and type of building Bridge and same heat source (fire starter blocs) Fire in a farm Fire in a farm which contained hay Vehicle fire Vehicle arson Many fires on the public path Wooden bridge fire Fire in an annex building of a pig farm Fire under a hay tunnel Vehicle fire Vehicle fire in a garage Fire of a small building on the public path Wooden bridge fire Rachel Robin Robin Robin Julien Julien

Relation Old case New case Insp

Table 2 Summary of similarities between different cases.

Aim of the relation

E. Erne, et al.

4.3. Knowledge conservation This ethnographic study shows that different methods still coexist within the forensic department in order to conserve knowledge gained by case investigations: some are institutional and other individual. The department manager sets institutional strategies, whereas investigators themselves set individual strategies. Two methods considered as institutional are the information-management systems (Zephyr and LIMS) that are used to archive cases. A third element contributes to knowledge conservation at the institutional level: written and illustrated synopses of interesting fire incidents that were investigated. These synopses are periodically prepared by a member of the fire-investigation specialist team and distributed by email to all investigators of the department. At the individual level, each investigator mentally conserves and structures his own knowledge. Explicitly, Julien said, “In fire investigation I remember all my cases very well. I do not know why, but they stay in my mind”. In addition to the mental memory, some investigators created physical and/or digital repositories of cases they investigated. For example, Marco and Aurélie constituted a paper repository that contains all the reports of the fire incidents they investigated and use it as a support for future report writing. Marco explained, “I like the option of having my folder behind me, with the reports that I wrote, and of having the possibility to read them and to turn pages without jumping from a computer screen to another” ThEse results illustrate the need for investigators to constitute their personal repositories outside of institutional tools in order to have rapid access to specific information. Marco indicated a preference for paper support, but it seems that the impossibility of having easy access to all the reports he wrote with Zpehyr or LIMS is also playing a role. The design of these databases requires accessing each case separately instead of having access to all of them at the same time. Julien, who was part of the fire investigation specialist team, made a paper repository that he calls the “Fire Bible”. This “Fire Bible” contained diverse information accumulated through years of practice and is considered by Julien as potentially useful for future fire investigations. He said, “Most elements of this Fire Bible are things that I use for most cases. But when you perform searches, you find things that you think could potentially be useful for a future case. I have also a lot of books about fire investigations or topics such as electricity.” The information contained in this “Fire Bible” has multiple origins: cases, fire investigation education, conferences, Internet, etc. Consequently, most of the knowledge is conserved in only the “Fire Bible” and is used almost exclusively by Julien, even if it is freely available to anyone within the department. 4.4. Knowledge acquisition and sharing On the institutional level, two strategies were identified. The first is the presentation of investigated cases; it occurs during the daily department meeting, every morning. Each investigator on duty has to attend the meeting during which all cases investigated the previous day are exposed and discussed, with the support of pictures. The second strategy is the synopses of cases previously mentioned; it is distributed to all investigators. Our interviews revealed that most investigators first considered that no institutional strategy was set in order to acquire the new knowledge. But, after we mentioned the two observed strategies, they admitted that indeed these enable them to acquire and share the knowledge. Laura 6

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

who are present”.

realized that the morning meeting was indeed a relevant way for continuous education for all investigators. She stated,

In addition to this, five investigators indicated during their interviews that within a few years, most of the experienced investigators would retire. They notified their concern about not having anymore such knowledge available to them, as they do not feel confident enough to complete the investigation of some cases by themselves. Aurélie stated,

“The morning meeting is a good way to educate the others. As for myself, I take as much information as possible, and if something is not clear, I discuss it with the investigators afterwards.” However, one of them pointed out the problem of the lack of attention during the meetings; André said,

“Julien will retire in few years. It will be a problem for me if no one replaces him because I need someone to help me with fire investigations.”

“My memory is essentially photographic and not oral. So if someone is only speaking, I will not acquire and remember the knowledge as well as if it is presented in pictures.”

Even if Aurélie has a seniority of 15 years and has already investigated, with support of an expert, between 50 and 100 cases, she felt that she is not skilled enough to investigate future cases by herself. Apart from oral transmission of knowledge, a few inspectors indicated they sometimes use books or the Internet to fill gaps in their knowledge. When the subject of personal education appeared in the discussion, all the investigators indicated that the best way to acquire new knowledge was to attend the scene of fire incidents and actively participate in their investigations. Aurélie stated,

Another investigator highlighted the problem of individuals that are not present at the meeting due to days off, part-time work or emergencies on the field. Concerning the synopses of cases, all interviewed investigators appreciate them, and some mentioned the advantage that the synopses could be read whenever they have time. Laura stated: “I think that it is always interesting and that it gives a good summaries of the cases.”

“My feeling is that the only way to acquire expertise is to go to the scene. At the office, cases can be shown to you but it is never the same as at the scene. Pictures are oriented and it is different from when you stand in the middle of the scene and everything is destroyed all around you. At the office, everything is prepared in order for you to understand the situation.”

But those summaries only concern about 15 cases a year, and there is the possibility that not all investigators read them. Regarding individual strategies deployed to acquire knowledge, two different situations were identified. The first one occurred during a fire investigation and was directly related to the case. The second appeared when an investigator wanted to enrich his knowledge, independently from a case. When facing a gap in knowledge during cases, investigators relied on their colleague’s knowledge. They mainly seek the help of more experienced investigators, or of colleagues who have knowledge in a particular field due to their previous education or previous cases. For instance, we observed that when a case involved a possible electrical fault, investigators often discussed the case with Florian as he was an electrician before becoming police officer. Laura confirmed this trend of using other’s skills in the department:

Due to operational constraints, generally only the investigator in charge of the case and one or two members of the fire-investigation specialist team go to the fire scenes. Consequently, some individuals will investigate only one or two fire scenes in a year, which significantly limits their opportunity to acquire knowledge in this manner. Four investigators also mentioned the use of the LIMS in order to gain knowledge. They examined pictures and the information about new cases added to the database, learning from fire-incident investigations they were not directly involved in. Robin explained,

“A lot of people have technical knowledge here, I will ask them to explain the situation”.

“Sometimes, information is given during the daily report, then I have a look at the case in more detail, especially in domain in which I have limited knowledge or that is more specific like electricity or explosions”.

It is interesting to note that even the most experienced investigator, Julien, relied on other’s knowledge, either within the forensic department, or within the Swiss fire investigators community. He stated,

Through participant observations, it was also noted that Julien and Marco systematically studied the pictures of every new fire incident that was investigated by the forensic department. In many cases, their interest was not solely related to case consultation in the database, and was completed by a discussion with the investigator who was in charge of the case. As details were often difficult to see on pictures of burn elements, this discussion allowed Julien and Marco to complete and nuance their opinion about the case and also to advise the investigator, for future fire incidents. This attitude contributed to reinforce their knowledge gained through fire investigation, and to share with others the experience they have accumulated.

“I have a good contact with some members of the Swiss fire investigation group who have a lot of technical knowledge. I can call them or send them an e-mail, and they will answer me. It’s a network that functions really well”. We observed that knowledge was mainly transmitted orally and directly through discussions. Knowledge exchanges could occur directly on the field, as many cases were investigated by more than one person (generally the investigator in charge of the case and a member of the fire investigation specialist team). It also took place at the office, with investigators discussing the case. Aurélie explained,

5. Discussion

“If I have questions, I go to Julien’s office to see if it is possible to determine, based on pictures, the cause of the fire”.

The findings of the ethnographic study support the hypothesis that, from a cognitive point of view, expertise has the same impact on individuals performing fire investigations as it does in other domains. The results we obtained during the interviews and our observations show that investigators consider the individuals who (within their department) have substantial expertise as experts in fire investigation. This status is independent from the internal organisation of the forensic department, particularly concerning the composition of the fire-investigation specialist team. Many investigators often require help from these experts, This shows that expertise is held as a central element in fire-investigation case solving.

And Marco added, “I share my office with Julien and people come to ask him questions, so we discuss these together”. This illustrates that design of office space also plays a role, in terms of knowledge sharing and dissemination. During the interview, André mentioned the problem caused by the absence of someone whose knowledge is required: “This human sharing of knowledge is clearly determined by individuals 7

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

the forensic department use this strategy. Another method for gaining expertise is to go on the field when investigators are not in charge of the case. For the investigators interviewed, it is the only way to acquire “real” expertise, as they have to realise the entire reflexion process without facilitation. This is obviously one of the best strategies, as they acquire direct experience and can reason by themselves [15,16]. Furthermore, this strategy enables the investigator do develop their skills through practice and to gain tacit knowledge as far as contributory expertise [11,13]. However, the operational constraints and sometimes the size of the scene do not allow for many extra investigators on the scene, rendering the accumulation of experience difficult. Generally, only two investigators go to a fire scene. All those elements contribute to knowledge transfer; however, due to the limitations discussed in this article, these elements do not overcome the long training required to accumulate the amount of knowledge or expertise required for novice investigators by themselves to investigate fire scenes. The concerns exposed by some investigators with 15 years of experience regarding the upcoming retirement of a colleague are symptomatic of this problem. These facts lead us to conclude that trying to increase each investigator’s personal expertise in order to transfer fire-investigation knowledge to prevent loss in the case of an individual’s departure is not a suitable solution. Management strategies, such as supervision and community of practice, are not the answer to this problem. Furthermore, these two strategies are still informally used - through the support of experts and discussions between colleagues - in this department and did not give sufficient results. This is why we propose to use a strategy that will enable the use individual’s expertise, without requiring their previous personal acquisition. This is precisely what case-based reasoning provides. In order to support case solving, it simulates expert reasoning by using a collection of past cases that could be retrieved through similarities. This means that another’s expertise could be used, at any time, without a previous acquisition of this knowledge. It is a way to artificially extend each person’s expertise. The degree of automation of an interaction with the user varies [33]. To solve a case, ire investigation is a domain in which much information, such as burn-pattern analysis and image analysis as testimony have to be combined with value judgments. They are ill-structured cases as defined by Jonassen [38]. At this step of the research, these observations lead to the definition that the best knowledge-management strategy is a low automation one. We then propose to use a case library as a fire-investigation knowledge management strategy. The construction of a case library will create a collective knowledge memory by externalising, preserving, and sharing everyone’s knowledge. Each investigator could then use the expertise of another, by searching for useful cases by comparing similitudes in the memory. The ethnographic study highlighted the fact that the comparison of related cases is a basic strategy used by everyone in the forensic department. But difficulties lie in the accessibility to the cases. In our opinion, the use of a case library could greatly facilitate access to them. This could tremendously impair the critical human factor currently ruling fire investigation in the forensic department and could guarantee a long-term conservation and sharing of knowledge. As the core of such strategies is to reproduce, at least partially, previous work, if a mistake is not identified, there is risk of perpetuating this mistake. This is why such a system is only for case solving and not to replace expert knowledge and critical thinking. We think that our study highlights notable benefits that can be derived from a reflective and critical approach, carried out in the form of an ethnographic study. The fact that the main author of this study was part of the forensic department under investigation was definitely effective in preventing misinterpretations of the observations. However, this might have hindered an attitude of detachment and of critical reflection. It is clear to us that the results of the ethnographic study, although useful for the specific forensic department that was considered,

Cognitive differences were also highlighted by the ethnographic study. To explain precisely his reasoning, Robin (a novice investigator) was able to remember all the situations where he studied and used the similarities. In contrast, Julien (the fire investigation specialist) had trouble remembering all such situations and was unable to cite only the multiple examples highlighted by observations. This proves that he really understood and fully used analysis of the similarities between cases. According to the literature review, this difference results from expertise variance and the consequential change to memory structure and cognition. The automatic and unconscious reflexion done by Julien is probably due to a combination of a complex and highly organised memory structure and the use of similarities at different levels of abstraction [4]. This finding confirms the need to externalise the knowledge of these individuals in order to make it available to others. The ethnographic study highlighted the pivotal role of similarities with past situations in the reasoning scheme that is applied for the investigations of fire incidents. In almost every case investigated during the time of this study, similarities were used by the investigator, or by a more experienced agent, to support case solving. The similarities were established to connect the fire incident under investigation with pastsolved cases in order to gather useful information. This role of similarities is fundamental: the finding of relevant past cases is a good predictor of the chance of solving a fire incident. Although rather schematic, this assertion highlights the fundamental necessity of having access to a large repository of cases. As we have seen, the memory of past cases is distributed and inherent in the investigators of the department. There is a plurality of individual and internalised memories. The ethnographic study brought to light some initiatives for externalising this knowledge, for instance, through the constitution of physical repositories. But these initiatives are neither systematic, nor uniform in their structure, hence they remain with the individuals who created them. The information- management systems that exist are mainly oriented towards the classification and storage of administrative information or forensic data. They are of limited utility in structuring, preserving, and sharing knowledge of the problem solving approaches of underlying investigation and reconstruction processes. Strategies are set to transfer knowledge hence broaden an individual’s expertise. The morning report, consisting in an oral presentation of each new case, is a good strategy, but the absence of some investigators due to operational constraints and the lack of attention create inequality in knowledge transmission. Even if all cases are presented, all investigators do not acquire their knowledge, as receiver’s motivation is the main limitation to this knowledge transfer [37]. Statements of interviewed investigators enable us to realise that the added value of case presentation is highly variable between individuals. A synopsis of a case is distributed to every member of the department but it only concerns a minority of cases (about 15 per year). All investigators receive the same information but, unlike with morning-report technique, the amount of knowledge transferred is very low. Consequently, institutional strategies are insufficient for transferring knowledge, because their efficiency is limited either by the number of cases concerned or by the number of investigators who benefit from this transfer. In parallel, these two strategies are based mainly on the transfer of explicit knowledge. As the acquisition of tacit knowledge is the main condition for developing expertise in a domain [11–13], the lack of efficiency of these strategies is easily understandable. Some inspectors developed individual strategies to accumulate more knowledge, but not all of them did so. The main methodology used is to study cases investigated by others within the department, either through the consultation of these cases in LIMS or through discussions with the investigators. By doing so, the individuals gain knowledge regarding the cases that can be used in the future. Such a process is valuable, as the investigator proceeds to a “second investigation” hence acquires practical expertise and tacit knowledge. But this is time consuming and requires motivation. This is probably why only some investigators of 8

Science & Justice xxx (xxxx) xxx–xxx

E. Erne, et al.

might not prevail in other services, and that generalization would be inappropriate. Therefore, we call for similar studies to be carried out in other institutions, in order to corroborate or amend findings from our study.

Homicide, Oxford University Press, Oxford, New York, 2003. [18] O. Ozkan, F. Dogan, Cognitive strategies of analogical reasoning in design: differences between expert and novice designers, Des. Stud. 34 (2013) 161–192, https:// doi.org/10.1016/j.destud.2012.11.006. [19] T. Gavrilova, T. Andreeva, Knowledge elicitation techniques in a knowledge management context, J. Knowl. Manag. 16 (2012) 523–537, https://doi.org/10.1108/ 13673271211246112. [20] S. Köpsén, S. Nyström, The practice of supervision for professional learning: the example of future forensic specialists, Stud. Contin. Educ. 37 (2015) 30–46, https:// doi.org/10.1080/0158037X.2014.967343. [21] M.S. Ackerman, J. Dachtera, V. Pipek, V. Wulf, Sharing knowledge and expertise: the CSCW view of knowledge management, Comput. Support. Coop. Work CSCW. 22 (2013) 531–573, https://doi.org/10.1007/s10606-013-9192-8. [22] J.-C. Chiem, T. Van Durme, F. Vandendorpe, O. Schmitz, N. Speybroeck, S. Ces, J. Macq, Expert knowledge elicitation using computer simulation: the organization of frail elderly case management as an illustration, J. Eval. Clin. Pract. 20 (2014) 534–543, https://doi.org/10.1111/jep.12101. [23] S. Doak, D. Assimakopoulos, How do forensic scientists learn to become competent in casework reporting in practice: a theoretical and empirical approach, Forensic Sci. Int. 167 (2007) 201–206, https://doi.org/10.1016/j.forsciint.2006.06.063. [24] E. Casey, Reinforcing the Scientific Method in Digital Investigations using a CaseBased Reasoning (CBR) System, University College Dublin, Dublin, 2013. [25] E. Roldan Reyes, S. Negny, G. Cortes Robles, J.M. Le Lann, Improvement of online adaptation knowledge acquisition and reuse in case-based reasoning: application to process engineering design, Eng. Appl. Artif. Intell. 41 (2015) 1–16, https://doi. org/10.1016/j.engappai.2015.01.015. [26] R.R. Hoffman, The problem of extracting the knowledge of experts from the perspective of experimental psychology, AI Mag. 8 (1987) 53–67. [27] A. Aamodt, E. Plaza, Case-based reasoning: foundational issues, methodological variations, and system approaches, AI Commun. 7 (1994) 39–59. [28] R.C. Schank, R.P. Abelson, Scripts, Plans, Goals, and Understanding: An Inquiry into Human Knowledge Structures, 1 edition, Psychology Press, Hillsdale, NJ, 1977. [29] J.L. Kolodner, J.N. Owensby, M. Guzdial, Case-based learning aids, in: Handb. Res. Educ. Commun. Technol. Proj. Assoc. Educ. Commun. Technol. 2nd Ed, 2003: pp. 829–861. [30] K.M. Kulasegaram, Z. Chaudhary, N. Woods, K. Dore, A. Neville, G. Norman, Contexts, concepts and cognition: principles for the transfer of basic science knowledge, Med. Educ. 51 (2017) 184–195, https://doi.org/10.1111/medu.13145. [31] A.A. Tawfik, A. Gill, M. Hogan, C.S. York, C.W. Keene, How novices use expert case libraries for problem solving, Technol. Knowl. Learn. (2017) 1–18, https://doi.org/ 10.1007/s10758-017-9324-1. [32] A.A. Tawfik, Do cases teach themselves? A comparison of case library prompts in supporting problem-solving during argumentation, J. Comput. High. Educ. 29 (2017) 267–285, https://doi.org/10.1007/s12528-017-9136-2. [33] J.L. Kolodner, Case-based Reasoning, Morgan Kaufmann Publishers, 1993. [34] S. Beaud, F. Weber, Guide de l’enquête de terrain, La découverte, Paris, 2010. [35] D. Wyatt, D. Wilson-Kovacs, Understanding crime scene examination through en ethnographic lens, WIREs Forensic Sci. (2019), https://doi.org/10.1002/wfs2. 1357. [36] S.L. Schensul, J.J. Schensul, M.D. LeCompte, M.D.L. Ph.D M.A., Essential Ethnographic Methods: Observations, Interviews, and Questionnaires, Rowman Altamira, 1999. [37] R.G. Schafermeyer, R.R. Hoffman, Using knowledge libraries to transfer expert knowledge, IEEE Intell. Syst. 31 (2016) 89–93, https://doi.org/10.1109/MIS. 2016.36. [38] D.H. Jonassen, Instructional design models for well-structured and III-structured problem-solving learning outcomes, Educ. Technol. Res. Dev. 45 (1997) 65–94, https://doi.org/10.1007/BF02299613.

Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References [1] O. Ribaux, Police scientifique le renseignement par la trace, Presses polytechniques et universitaires romandes, Lausanne, 2014. [2] E. Bruenisholz, O. Delémont, O. Ribaux, L. Wilson-Wilde, Repetitive deliberate fires: development and validation of a methodology to detect series, Forensic Sci. Int. 277 (2017) 148–160, https://doi.org/10.1016/j.forsciint.2017.06.009. [3] A. Waser, Méthodes et moyens d’analyse et de suivi des incendies non délibérés, Mémoire intermédiaire de thèse Université de Lausanne - Ecole des sciences criminelles, 2010. [4] D. Hofstader, E. Sander, L’Analogie coeur de pensée, Odile Jacob, Paris, 2013. [5] M.T.H. Chi, P.J. Feltovich, R. Glaser, Categorization and representation of physics problems by experts and novices, Cogn. Sci. 5 (1981) 121–152, https://doi.org/10. 1207/s15516709cog0502_2. [6] G. Marchant, J. Robinson, U. Anderson, M. Schadewald, Analogical transfer and expertise in legal reasoning, Organ. Behav. Hum. Decis. Process. 48 (1991) 272–290, https://doi.org/10.1016/0749-5978(91)90015-L. [7] E.J. Johnson, Expertise and decision under uncertainty: performance and process, in: T.H.C. Michelene, R. Glaser, J.F. Marshall (Eds.), Nat. Expert. Lawrence Erlbaum Associates, Hillsdale, New Jersey, 1988, pp. 209–228. [8] H.G. Schmidt, G. Norman, H. Boshuizen, A cognitive perspective on medical expertise - theory and implications, Acad. Med. 65 (1990) 611–621, https://doi.org/ 10.1097/00001888-199010000-00001. [9] M.T.H. Chi, Two Approaches to the Study of Experts’ Characteristics The Cambridge Handbook of Expertise and Expert Performance, in: Camb. Handb. Expert. Expert Perform., Cambridge University Press, 2006. [10] J. Shanteau, How much information does an expert use? Is it relevant? Acta Psychol. (Amst.) 81 (1992) 75–86, https://doi.org/10.1016/0001-6918(92) 90012-3. [11] S. Köpsén, S. Nyström, Learning in practice for becoming a professional forensic expert, Forensic Sci. Int. 222 (2012) 208–215, https://doi.org/10.1016/j.forsciint. 2012.05.026. [12] M. Eraut, Non-formal learning and tacit knowledge in professional work, Br. J. Educ. Psychol. 70 (2000) 113–136, https://doi.org/10.1348/000709900158001. [13] H. Collins, R. Evans, Rethinking Expertise, University of Chicago Press, London, 2008. [14] J. Van Maanen, Observations on the making of policemen, Hum. Organ. 32 (1973) 407–418, https://doi.org/10.17730/humo.32.4.13h7x81187mh8km8. [15] P.A.J. Waddington, Police (canteen) sub-culture. An appreciation, Br. J. Criminol. 39 (1999) 287–309, https://doi.org/10.1093/bjc/39.2.287. [16] D. Wyatt, Practising crime scene investigation: trace and contamination in routine work, Polic. Soc. 24 (2014) 443–458, https://doi.org/10.1080/10439463.2013. 868460. [17] M. Innes, Investigating Murder: Detective Work and the Police Response to Criminal

9